Spectral Computed Tomography Development

Abstract

The goal of this thesis was to improve spectral CT. This was done through software, scanning, and camera development. The intention was to make a stable and robust system that will enable clinical validation of spectral CT concepts.

A spectral CT scanner for small animals was designed and built by the MARS team. An important component of spectral CT is the detector. The MARS project uses the Medipix ASICs, which are a development of energy discriminating photon counting detector. Several iterations of this scanner underwent development in response to evolving needs of the team. A GUI was developed to enable robust high quality CT scanning of spectral CT images. The GUI was created in several iterations, including the latest MARS GUI v5.0. It can run continuous motion circular, start-stop circular and start-stop helical scans. It sends DICOM files to a PACS server, and can run custom scripts for calibration. The Medipix3 is the newest family of Medipix chips. Several Medipix ASICs were characterised: the MXR, Medipix3.0, Medipix3.1 and Medipix3RX. These were equalised and used for imaging purposes. The equalisation process for all four significant versions of the chips were automated in software. The Medipix chips were validated with energy sources, and a new technique was developed for energy calibration. The thesis makes contributions to original research by 9 published papers and 1 paper in preparation. This work is significant because it demonstrates that spectral CT is in a state where it can easily be used by researchers. The development of the GUI allows for scans to be done and stored professionally. Spectral data sets have been acquired with the Medipix chips to demonstrate the ability to perform material decomposition. With these accomplishments, it should be clear to any observers that spectral CT is a useable technology with potential for molecular and functional imaging of radiographical contrast agents.